Every few years, someone in the organisation decides the website needs “a complete overhaul.” New branding, new layout, new everything. It feels decisive. It feels like leadership. It is, almost without exception, one of the most reliably destructive decisions a marketing leader can make.
The instinct to tear it down and start over is seductive because it skips the hard work of understanding what’s actually wrong. It replaces diagnosis with demolition. And it nearly always makes things worse.
Table of contents
Open Table of contents
The $62 million weather forecast
In October 2025, Australia’s Bureau of Meteorology (known affectionately as “the Bom”) launched a redesign of its weather website — one of the most-visited sites in the country, with 2.6 billion views a year. The previous design hadn’t been updated in over a decade.
Within hours of launch, the hashtag #changeitback was trending.
Farmers couldn’t find the GPS coordinate search they relied on to check conditions for specific paddocks. Emergency workers lost the radar colour scheme they’d spent years learning to read. The site’s acting head was forced to publicly apologise within a week.
Then it got worse. It emerged the redesign had cost A$96.5 million — twenty times the originally stated A$4.1 million. Barnaby Joyce, a member of the Nationals, summed up the national mood: “We spent $96m to put a B at the end of the Bom site. It’s now bomb, it’s hopeless.”
Psychologist and neuroscientist Joel Pearson explained the outrage to the BBC:
“It’s the government IT project equivalent of ordering a renovation, discovering the contractor has made your house less functional, and then learning they charged you for a mansion.”
If you have ever been tempted to “tear it down and start over” with your website, this BBC article on the Bom redesign is a must-read. It serves as a stark reminder that even with the best intentions, a complete overhaul can backfire spectacularly when it loses sight of user needs.
Why redesigns destroy value
The problem with redesigns isn’t that they change things. It’s that they change everything at once, with no way to measure what helped and what hurt.
When users visit your site regularly, they build what neuroscientists call “procedural memory” — the unconscious patterns that let them navigate without thinking. It’s the same kind of muscle memory as touch-typing or driving a familiar route home.
A redesign obliterates all of that overnight. As Pearson explained, the Bom redesign was “scrambling the neurological shortcuts that people had spent a decade building.” He compared it to rearranging all the furniture in someone’s house and expecting them to navigate it in the dark without stubbing their toe — except the “furniture” in this case determines whether a farmer moves their livestock before the flood arrives.
This is why redesigns almost always tank conversions in the short term. And why most never recover. You can’t A/B test your way out of a decision you’ve already shipped to 100% of users.
The redesign death spiral
Most redesigns follow a depressingly predictable pattern:
- The current website “isn’t performing” (though no one can say exactly why)
- An agency pitches a beautiful new design backed by a flashy portfolio and creative awards
- Months of development, internal politics, and scope creep follow
- Launch day arrives and conversions drop
- Everyone blames “the transition period” and waits for things to improve
- They don’t
- Two years later, someone proposes a new redesign
This is exactly the cycle described in How to tell if your expensive new website is useless, which explains why big-budget redesigns routinely fail to improve conversions. The core problem: agencies build visually stunning sites without ever measuring whether they actually drive business results. Conversion is a website’s purpose for existing, yet it’s treated as an afterthought — something to “optimise later,” after the real work of choosing fonts and hero images is done.
When redesigns work (and how)
Sometimes a redesign is genuinely necessary. Legacy platforms, security vulnerabilities, fundamental mobile responsiveness issues — there are legitimate reasons to rebuild. But successful redesigns are approached strategically, not emotionally.
The RealEstateU Redesign Win Report is a good example of how to get it right. After a rebrand caused a 30% drop in sales overnight, the team didn’t just revert to the old design and hope for the best. They ran user research, discovered that trust had been destroyed — users literally said the new site “seems like a scam” — and systematically rebuilt based on evidence.
The result: conversions exceeded the original site, with trustworthiness scores outperforming every competitor. Likelihood of users exploring further increased by 140%.
The difference between this and the Bom? Research-driven redesign vs. opinion-driven redesign. One starts with the question “what do our users actually need?” The other starts with “what would look cool?”
Your words are doing the heavy lifting
There’s a page by Justin Jackson that we keep coming back to. It’s just words on a white background. No animations. No hero images. No parallax scrolling. And it’s one of the most persuasive pages on the internet, translated into more than 35 languages.
As Jackson puts it:
“I wrote these words, and you’re reading them: that’s magical.”
It’s a powerful demonstration that you don’t need complex animations or heavy graphics to be persuasive. In fact, most of the time, your words are doing the heavy lifting. As Jackson argues elsewhere on the page, “web design should be about words. Words don’t come after the design is done. Words are the beginning, the core, the focus.”
The next time an agency shows you a Figma mockup filled with lorem ipsum, remember this. If they haven’t started with the words, they haven’t started.
What to do instead of a redesign
If your website isn’t converting, the answer is almost never a redesign. It’s optimisation.
- Measure first. Before changing anything, understand what’s actually happening. Where do users drop off? What are they clicking? What are they ignoring? If you can’t answer these questions, a redesign won’t help — you’ll just be rearranging deck chairs.
- Test small changes. Headlines, calls to action, page structure, social proof — these are the levers that move conversion rates. Test them individually so you know what works.
- Iterate continuously. The most successful digital companies don’t do redesigns. They make hundreds of small, measured changes over time. Amazon’s website today doesn’t look dramatically different from five years ago, but it converts better every year.
- Preserve user habits. Your existing users have learned how to use your site. That knowledge is valuable. Changes should enhance their experience, not reset it.
Stop redesigning. Start optimising.
A website redesign feels productive. It generates excitement, fills project boards, and keeps agencies busy. But it’s almost always a substitute for the harder, less glamorous work of actually understanding your customers and systematically improving their experience.
The next time someone in the room proposes tearing it all down and starting over, remember the Bom. A$96.5 million. Twenty times over budget. And a nation of angry farmers who just wanted to check the weather.